Background: although written from the point of view of Wittgenstein's logic of language (the meaning of Wittgenstein's expression 'logic of language' in my jargon), this paper should not be taken to represent Wittgenstein's own views about the philosophy of science.
M. O'C. Drury's Philosophy of Science
Outline of this page ...
- A Scientific Theory attempts to be a More Satisfying Version of Reality
- Why Theories Must not be Regarded as Facts
- What is Scientific Explanation?
- Is Seeing Theory-laden? Is Seeing Concept-laden?
- Explanation does not Naturally come to an End, but it must be recognized that Explanation does come to an End
- Seeing that some facts will always be unexplained (Philosophical clarity)
- The Value of Maintaining the Distinction between Theories and Facts
A Scientific Theory attempts to be a More Satisfying Version of Reality
I want to return now to Drury's more general remarks about theories and facts.
We have in immediate experience our one sole contact with reality, and everywhere this immediate experience cries aloud that it is incomplete and fragmentary. (M. O'C. Drury, The Danger of Words (1973) p. 109)
What did Drury mean by 'our one sole contact with reality' -- for "immediate experience" is not the only thing we call 'contact with reality' (if we call anything that, which it seems we do), unless Drury is contrasting 'immediate' with 'mediated' and by 'contact' means 'direct contact', in which case he is saying: "we feel our direct experience of reality to be incomplete and fragmentary".
Then we go on to construct in imagination the conception of an experience which would be more adequate, more satisfying. (ibid.)
By 'conception' I take Drury to mean his "models, pictures, maps" (ibid. p. 99-100). For example, someone observes the movements of the moon, the planets, and the sun in the sky, but sees no interrelated pattern of movement. He may see individual patterns, but not a system. What our theories -- geocentric and heliocentric -- do is to relate these individual patterns to one another, e.g. into a the heliocentric theory's solar system. And these models or pictures or maps, these mediate conceptions of the heavens, these systems -- some people feel to be more adequate or satisfying than the "incomplete and fragmentary" experience they have when they observe the sky.
North and south, Up and Down
"That is what every scientific hypothesis, apart from its practical usefulness, attempts to be" (ibid. p. 109) -- i.e. a more "adequate", more satisfying "version of reality" than the facts on which the theory is based are by themselves alone.
Maybe this section should be titled: "A scientific theory attempts to be a more satisfying version than reality".
And is this why the picture of the solar system is banged into the head of every schoolchild? "It explains to them where the astronauts and spaceships are." Well, it gives them a picture to relate that to (but a picture may mislead as well as make clearer | Pictures that don't misrepresent yet do mislead); and isn't such a picture one of the things we call an 'explanation'?
But why is the picture presented to the children as if it were a fact? For obviously that is not a grammatical requirement. Children are e.g. frequently told that we can turn the Map of the World upside down -- i.e. that 'north' and 'south' do not mean 'up' and 'down'. But we can imagine a people who insisted that this was precisely what those words do mean, and who treated the "right end up" Map of the World as a fact. They would be more like us than we would like to accept -- when we treat the heliocentric model of the solar system as if it were reality as it is in itself.
And in so far as such a process of inference does bring a greater sense of unity into our experience, it is so far legitimate. (ibid.)
A 'system' makes a unity of individual facts by relating them -- by imagining them into a single picture. But what did Drury mean by "such a process of inference"? Drury meant that the model or picture or map we construct must be consistent with the known facts (For a theory to be 'scientific', it must be justified in this way) -- and this can be called an 'inference'.
Why Theories Must not be Regarded as Facts
What is not legitimate is to think that the process of inference is at an end and the ideal is now reached. (DW p. 109)
... the logical status of a scientific hypothesis. That it is always a transitory, incomplete affair. Never finished, final, factual. Every scientific hypothesis is always at the mercy of new evidence and may require indefinite modification in the light of this evidence.
... Considering the vast complexity of the matrix of nature, isn't it certain that there is still much evidence, lying before our eyes and beneath our hands, which we have failed to notice as yet?
[There is] a universal human tendency to take transitory concepts as final and absolute. (ibid. p. 110)
There are two elements here: new evidence and new conceptions of old evidence. 'Theories' are those conceptions that must be consistent with evidence -- both that at hand and that which may yet be discovered. So they are "never finished, final, factual" (That 'factual' means: to be treated as a fact).
We may imagine e.g. (and perhaps quite ahistorically) that once upon a time the flat earth and domed heaven was the only picture of the universe at hand -- that human beings regarded that picture as a fact. They did not even conceive the possibility of another picture. Then the geocentric picture was imagined, later the heliocentric ... and wouldn't it be foolish to suppose that this is the end of it. -- Even though we admit that we assign no sense to 'absolute motion', nor to the claim that 'all the evidence must be in'. And that the limits of picture-making are indeterminate; so that, every scientific hypothesis is also always at the mercy of some human being's imagination. (A limit of science is the logically possible, the "what can be described", the "conceivable". But what 'can' be described is what is described. And who would presume to set any limit to what may be described some day.)
"The giving of explanations comes to an end. If it didn't, there wouldn't be explanations." (cf. Philosophical Investigations § 485; this is of course a grammatical remark; it is not a remark about human limitations.) Drury wrote that "every spatial and temporal picture goes to pieces completely at its edges" (DW p. 109). E.g.:
Astronomers are rightly interested in finding an explanation for the remarkable "red shift" in the spectral lines of very distant nebulae. The generally accepted explanation is that it is a manifestation of the Doppler effect (which is familiar enough to us in the way the pitch of a rapidly moving whistle of a train changes as the train either approaches or recedes from us).
So it is thought that these nebulae are receding from us at prodigious speeds. Now this is a possible scientific explanation, and in one sense it makes the phenomena clear. But then we at once want to ask, "Why are these nebulae receding at such speeds?"
And this shows us that ultimately we will have to accept some facts [Note 10] as unexplained, and say, "Well, that is just how it is".
So then there would be nothing illogical in saying of the shift in the spectral lines, "That is just how the spectra of distant nebulae are", and we are not forced to give any explanation. (DW p. xi-xii)
Note the words 'the generally accepted explanation'; because at school we are told that it is the explanation. That is, at school this theory is presented to us as if it were a statement of fact.
What is Scientific Explanation?
Is an explanation in science a statement of theory or a statement of fact? Because the practice of looking for facts that verify propositions does not entail the construction of theories (i.e. of Drury's "pictures, models, maps"), then if we are to give a unique sense to the word 'science', then a 'scientific explanation' is never a statement of fact. But, then, in what sense is a theoretical statement an 'explanation'? because there are also explanations, e.g. of meaning, in logic and these are not theoretical.
I might say: "I am happy to see the sun rise and set, and I ask for nothing more." But suppose I did want to ask what the relation of this rising and setting sun to the earth was -- wouldn't the heliocentric and geocentric ways of looking at this relationship -- aren't these two models examples of what we call a 'scientific explanation'?
Difference between description and explanation
There is a distinction between describing the facts -- i.e. saying what they are -- and explaining the facts -- i.e. saying why they are what they are --. This is a divergence between the grammars of 'description' and 'explanation', where they don't overlap with one another. We on the earth see the sun rise in the sky in the morning and set in the sky in the evening. That is a description. The sun circles the earth or the earth spins while the sun stands relatively still; these are explanations, ways of looking at the facts (They are models of relative motion). And they do answer the question people presented with a flat earth must eventually have asked: where does the sun go after it sets in the evening -- and how does it get back to the point from where it rises in the morning?
But what of the explanation that the earth is round? (Here we must recall the days, not so very long ago, when 'The earth is round' was not a verifiable proposition; it was theoretical.) If the earth doesn't rest on anything, then why doesn't it fall down? We answer, because 'up' and 'down' are undefined in this context. Objects fall down to earth, but in which direction should the earth fall -- i.e. where is 'down' in this picture? This is a grammatical, not a theoretical explanation; it is an explanation that belongs to logic, not to science (although, of course, it arose within the context of a scientific theory).
Is Seeing Theory-laden? Is Seeing Concept-laden?
Usually when we speak of 'making observations' we mean looking and seeing. And looking and seeing is sense perceiving. But, we are reminded: This is not just a simple matter of opening our eyes. If e.g. I look around the room, I see a table and a bookshelf, a window and a tree. And not just "a tree", but a red maple tree. It seems, in this example anyway, that I don't as it were see bare reality -- but that I see in our language. Is seeing, then, concept laden? But were we to say that "all seeing is concept laden" -- or that "all seeing is theory laden", then we would be either stating a rule of grammar or talking nonsense.
The expression "the conceived facts" is used in a letter from Albert Einstein to Max Born. Maybe, for there is no way to know (i.e. this is metaphysical speculation), there are raw percepts, a raw reality, but man knows percepts only as mediated by his concepts. There are no facts before man conceives them ("This is of course a grammatical remark").
N.R. Hanson wrote: "Seeing a bird in the sky involves seeing that it will not suddenly do vertical snap rolls ... This is knowledge: it is knowing what kind of thing" the word 'bird' denotes. [Note 11] There are many clarifications to make here.
First, 'Birds do not suddenly do vertical snap rolls' is not a statement of fact; it is a generality, and so a theory about birds. It allowed Hanson to predict what any given bird would do, or not do; and a prediction is not a statement of what we know: it is not knowledge (and so not a fact either -- i.e. the grammars of 'fact' and 'knowledge' are connected).
Second, this is related to the case of Waismann's cat that suddenly grows to the size of a house and barks -- is it still to be called a 'cat'? But what do we need to know in order to answer Waismann's question? The insight here is to see that there is no grammatical rule for whether Waismann's cat is still to be called a 'cat'. And if a bird were to "suddenly do vertical snap rolls", we would not know if it were still to be called a 'bird' -- because there is no rule in our grammar to cover this particular case either. [Note 12]
And if I say "But a bird can't ...", then my thinking is "theory laden".
But don't we sometimes say: "What was that? At first I thought it was a bird, but now ..."? Of course we do; but normally what we mean is that we did not get a good look at the thing, and that what we did see did not go as we expected it to. So we do have expectations about what birds will and won't do? Of course we have, and Hanson gave an example of one. But is it natural to say that these expectations "belong to the phenomenon of seeing" -- i.e. to say that "seeing is theory laden"?
Is Seeing Fact-laden?
Macbeth's "tomorrow and tomorrow and tomorrow" assumes the future constancy of nature. Do we only believe it's to "the last syllable of recorded time"? or do we know that? It seems presumptuous to say we know that; and it is logically possibly false. On the other hand, if "We don't know whether the sun will rise tomorrow" is not a false proposition, then is any prediction a true proposition?
Do we know what is well-established by experience? Is this only a question of how we choose to define the word 'know'? For many years ago I have been thinking about this, and I haven't a different answer now from my first answer (and I still am not satisfied with the old one, although it's correct that 'knowledge' ≠ 'conviction', that 'know' and 'convinced' are different concepts).
"Are facts fact-laden?" Only if past events are used to predict future events and to explain present events -- and past events are used that way (PI § 481). Don't I know whether we had coffee this morning; do I only think I remember that we did? (Strange expression: "based on past experience" -- in contrast to what kind of experience?) We use the word 'know' in all sorts of (different) ways -- There is no general definition (essence, common nature).
How do we define the word 'bird'? "What kind of bird are you: you can't even fly!" -- "What kind of bird are you: you can't even swim!" Sasha and the duck say to one another in Peter and the Wolf. We begin by pointing, as we begin with the names of flowers, only later learning those flowers' culture, whether moist soil, full-sun, things like this. Question: and those things we learn, do they become part of the definition of e.g. the word 'marigold'?
Where exactly are the limits between verbal and real definitions, between logic and hypotheses? (And so those two concepts are more than a bit fluid, tools that may melt in your hands. On the other hand, what is the purpose of definition in philosophy?)
A Theoretical Statement can be made a Rule of Grammar for a Special Purpose
Suppose we had asked Hanson: if you were to see a creature "suddenly do a vertical snap roll", and then land, and you examined it, and it was in all other respects (according to his theory) a bird -- would you call it a 'bird'? Hanson could have made any rule he liked here, decided how he wanted to define the word 'bird' for his special purposes, choosing which criterion he wanted to make defining. 'Birds do not suddenly do vertical snap rolls' could be made a rule for applying the word 'bird', as might have been natural for Hanson, who was a airplane pilot.
Why shouldn't a theoretical statement be made a rule of grammar? Doesn't this actually happen in science?
In the beginning, the word 'fermentation' named the bubbling up that occurs as e.g. grape must becomes wine. This process starts off violently, and the must is said to be "in ferment". Many centuries later it was discovered (and generalized into a theory) that it is the activity of microbes that makes the must bubble up: yeast in the absence of oxygen change grape sugars into carbon dioxide and ethyl alcohol; the carbon dioxide "bubbles up".
But yeast were also found to change grape sugar into carbon dioxide, though not alcohol, in the presence of oxygen. And in the original use of the word the latter process would also be called 'fermentation', since that word was not originally defined theoretically, but "naively".
Nonetheless, 'fermentation' is now defined as an anaerobic process. Why? I imagine because vintners pushed for the original research (They wanted to know how to prevent their wine from turning into vinegar), and there is no wine without alcohol. This is an example of a theoretical statement being made a rule of grammar for a special purpose.
Is a Generalization a Statement of Fact?
But aren't I talking unnaturally when I call general statements 'theoretical'? Is it a 'theory' that human beings get dizzy when they spin round and round? Choose for yourself (i.e. there does not seem to be a grammatical rule here). Do you want to call a prediction 'knowledge'? (This is the way philosophers think, namely metaphysically.)
But, then, am I rejecting Newton's Rule IV? No, -- or not entirely: I am just not calling a "proposition inferred by general induction" a 'fact'. A generalization may be "accurate" or inaccurate -- i.e. it may be well-founded or barely justified, or anything in between, but I am not willing to call even the most well-founded generalization "very nearly true"; because what cannot be true cannot be nearly true either (A grammatical remark). A statement of fact is true or false; a generalization is well-founded, etc., or disproved -- but never true; it is always subject to being disproved by a single counter-instance; and so I am calling it a 'theory'. To induce from A, B, C, ..., have died that All men are mortal is to make a generalization, a summary of the facts (the facts being that A died, B died, etc.); and so the statement 'All men are mortal' is a theory -- and however well-founded it is (and no theory is more well founded than this one), it remains subject to disproof. "You can trust your life to its predictive power" -- but it is still not a fact, even if we are satisfied that all men are mortal (cf. OC § 299). The question is whether the statement is certain, not whether we are certain.
As I remember, and I often remember things wrong, Alice said, "I didn't know that cats could smile." And the Duchess replies, "All can. Some do." (Alice in Wonderland, vi) But how do we know that the cats that do not smile can smile? (And what does it mean to say that they "can" if they never do?)
I want to maintain the grammatical distinction between 'statement of fact' and 'generalization'. But not because I want to put something else in the place of induced generalities (for I too want to "frame no hypotheses").
I want to maintain the distinction." But -- why? Possibly because it has never before occurred to me to make it -- and now I am astonished by my blindness to it. I had never thought about it before.
We should not treat a generalization as though, as if were an absolute. All birds do not fly and all birds to not swim (Peter and the Wolf). (A generalization is like a scientific theory in that it is always possible for further evidence to falsify it.)
A Rule of Grammar is not a Theory
Third (clarification with regard to N.R. Hanson), our normal way of talking does not contain any theory about birds, but only a rule for using the word 'bird' (cf. Zettel § 223). The word 'bird' is defined ostensively; that is how we learn to use that word; that is its grammar. "So what the things you learned to call 'birds' look like is the criterion for applying the word 'bird'?" 'Bird' is a fluid concept: we do not use a specific criterion, that can be stated in a general rule, to apply the word. We learn to use the word 'bird' in the presence of birds -- and that is as far as our instruction in the language goes. We could make a rule, out of Hanson's statement e.g., but that would not be a rule that belonged to the grammar of our everyday word 'bird'. 'Birds do not suddenly do vertical snap rolls' is not a rule of our common grammar. In that grammar there just is no rule.
Fourth, "Hanson wanted to make a point about the phenomenon of perception in science, not about grammar"? Nonetheless his remark lands us in grammatical questions. Because 'all seeing is theory laden' is either a rule of grammar or nonsense, in just the way that 'all objects have dimensions' is. (cf. OC § 35-36)
Sometimes, e.g. in science, seeing may be theory-laden. But it is a blunder when this happens. Because we do not want to look at the evidence with prejudiced eyes. Just as we want to draw careful distinctions between grammar and statements of fact, and between facts and theories. (There may also be some reason to say that some seeing is "grammar laden". But first, what will that mean? And second: Look and see.)
Explanation does not Naturally come to an End, but it must be recognized that Explanation does come to an End
I believe that "the generally accepted explanation" for the prodigious receding of the distant nebulae is the so-called Big Bang theory. This is interesting for being an example of a theory twice removed from the facts, if indeed the "red shift" is a fact (and I don't know whether it is). But the next questions are, of course: why was there this primeval unity, and why did it "explode"? (Now perhaps we will need theories three or four times removed from the facts.) So that, we never reach a final resting place in science: there is always another why. And there is nothing whatever to prevent these theories from being discarded to be replaced by others -- of a kind as yet unimagined (just as our ancestors perhaps could not imagine anything but a flat earth model).
Drury quotes TLP 6.371:
The whole modern conception of the world is founded on the illusion that the so-called laws of nature are the explanation[s] of natural phenomena. [tr. Pears & McGuinness]
Wittgenstein somewhere used the expression 'the laws of physics'. I think that for Drury's purpose here we need to underline the final 'the' and add: rather than the transitory "models, pictures, maps" of the sciences (or the scientists) that they are. [Note 13]
Seeing that some facts will always be unexplained (Philosophical clarity)
Philosophical clarity then arises when we see that behind every scientific construction there lies the inexplicable. (DW p. xii)
And 'inexplicable' is the word Drury meant. Whereas the phrase 'the as yet unexplained' is not what Drury meant: because, not only are there endless whys, there are also limited-only-by-human-imagination possible explanations, not just the one.
... we will have to accept some facts as unexplained, and say, "Well, that is just how it is."
Drury's point here is a philosophical point. [Note 14] But also: when scientists use words like 'fortuitous', 'de trop', 'random', 'accidental' or 'chance' -- they are characterizing their work, not reality; what they are saying is: we are not going to try to explain this, but will accept it as given. (On the other hand, further research does require -- logically require -- something further to research.)
The Value of Maintaining the Distinction between Theories and Facts
Drury made a distinction between facts on the one hand, and hypotheses or theories on the other. By 'hypothesis' Drury meant the "models, pictures, maps" that the scientist invents to summarize a complicated collection of facts. The hypothesis is a picture, a work of the imagination with no independent existence. The facts are what actually exists; they are the raw data from which the scientist selects to create his pictures. By 'facts' Drury simply meant what human beings actually observe (perceive) to exist: the stone with indentations versus the clothed-in-flesh dinosaur, the bird's chirping versus the sound wave, the kitchen table versus the atoms-in-the-void.
We have been brought up in the Darwinian tradition. Our popular books, our encyclopaedias, our natural history museums, have presented this one hypothesis as a fait accompli. It is an hypothesis that is now so familiar that it is mistaken for a fact. (DW p. 111-112) [But:]
A theory cannot become a fact, but it can become a dogma.
"It keeps wonder secure"
Every scientific hypothesis is a transitory and to some extent arbitrary affair [i.e. a question of selection and imagination]. It must never be allowed to solidify into a pseudo-fact. But why not? What harm is done? ... You ask what is the value of such scepticism, such agnosticism, such carping criticism? (ibid. p. 113-4)
The task of philosophy
Drury answered: "One value only. It keeps wonder secure." (ibid. p. 114 [Note 15]) Drury thought this -- as he understood Wittgenstein to say (The Blue Book p. 45: "The difficulty in philosophy is to say no more than we know"; PI Motto: "It is the nature of every advance that it appears much greater than it actually is") -- to be the task of philosophy (and so too of the philosophy of science):
To insist that people say only just as much as they really know; that, when as happens in every generation, new advances in knowledge are made, they are not taken to be more important than they really are. (DW p. 113-4)
Drury saw Socrates as having started philosophy off on this course.
Note 10: I would like to replace the words Drury wrote here with the word 'something'. Drury wrote 'some facts' -- and, in this context, those words may not make sense, given everything else Drury has said, because it is not a fact that "these nebulae are receding from us at prodigious speeds"; that is an hypothesis. [BACK]
In the context of "Facts are theory laden", there is Kant's "Percepts without concepts are blind", and Goethe's Das höchste wäre zu begreifen, daß alles Faktische schon Theorie ist ("The most important thing to remember is that all fact is already theory", a possible meaning we might give Goethe's statement is that any fact only exists as a fact within some theory -- or, way of looking at nature -- we have conceived or preconceived, rather than somehow (I don't know how) letting nature be what it is rather than in any way forcing it to be what we wish it to be).
What does Goethe mean by alles Faktische ist schon Theorie? That "all fact is already theory" / "Everything factual is already theoretical" / "All facts are merely theoretical constructs"? Question: but if everything is a theoretical, then nothing is? What, then, however, are frames of reference -- or is there "an absolute point of view"? Any fact is only a fact within some frame of reference ("theory", "way of looking at things", "relative to some reference point/s" or even "space frame") or other; none is absolute. [BACK]
But perhaps I want to object: "What matters here is not what we call the creatures, because that is merely a linguistic convention; -- no, what matters here is what the creatures really are." But what might I mean by this? A cat metamorphoses into a house -- and I want to know what this creature is? But I am not asking what name I am to call it -- no, I want to know what it is. It is a cat that metamorphosed into a house; that's what it is. In other words, I cannot think of any sense to give to my "objection". [BACK]
Note 13: Drury warns that theories are temporary constructions and, so, cautions against making scientific theories the foundations of one's philosophical thinking (about any subject, but especially about the riddles of life and death): "Hypotheses, as Kant said, are contraband in philosophy." (DW p. 110) [BACK]
Note 14: Why -- "it is there -- like our life" (OC § 559)? Perhaps Drury saw here some part what Wittgenstein meant by this remark. [BACK]
Note 15: In his "Letters to a Student of Philosophy" (p. 167) Drury has a quotation that appears to be the origin of the expressions "keeps wonder secure" and "then wonder comes in in the right place":
It is true, as the understanding says, that there is nothing to wonder at, but precisely for this reason is wonder secure -- because the understanding vouches for it. Let the understanding condemn what is transitory, let it clear the ground, then wonder comes in in the right place, in ground that is cleared, in the changed man. Everything pertaining to the first wonder the understanding can consume; let it do so, in order that enigmatically it may help one to wonder, for it is indeed enigmatical, since it conflicts with the judgement of the understanding concerning itself. (Kierkegaard, "Three Discourses on Imagined Occasions")
Site copyright © September 1998. Send Internet mail to Robert [Wesley] Angelo. Last revised: 3 August 2007 : 2007-08-03
The URL of this Web page: